Engineered nanomaterials (ENMs) can be used in cosmetics as UV blockers. For these products, the exposure-driven risk for humans and the environment is related to the ENM release during or after use, and thus to the original surface properties and aggregation state of the ENMs. Moreover, as the ENM dispersion in the product also affects the efficiency to screen UV rays, optimizing the formulation can lower the ENM load. Characterizing the ENM behaviour directly in a cosmetic formulation is thus crucial to better assess their risk and develop safer-by-design products. However, the complexity of such a multiphasic system limits in situ characterization using most common analytical tools. Here, we present a novel methodology based on two-dimensional X-ray absorption (2D-XRA) imaging to characterize the dispersion state of ENMs directly in a sunscreen product. Two commercial nano-TiO2 UV filters, displaying different surface coatings, were used to prepare contrasting sunscreen formulations at increasing ENM concentration. Cryogenic scanning transmission electron microscopy (Cryo-STEM) was also used for comparison to evaluate the advantages and limitations of both methods in this context. 2D-XRA proved to be a powerful and rapid technique to analyze both UV filter dispersion in the formulation and the overall product homogeneity. This was enabled by thresholding areas of contrasting ENM densities in the 2D-XRA image, which reflected ENM aggregates, fine ENM dispersions, or voids with a lower UV protection. Image analysis also allowed semi-quantitative evaluation of the relative area of each density range, and of the aggregate size in terms of projected area. In comparison, Cryo-STEM provided a larger magnification than 2D-XRA, which enabled visualisation and sizing of the ENM primary particles, plus the distinction of the emulsion oil and water phases thanks to EDX coupling, but with a smaller and less representative volume of analysis and a higher cost in time and energy. This work is a step forward in measuring ENM behavior in situ in a complex multiphasic matrix constituting a nano-enabled product. Such knowledge at the original stage of the product life cycle is crucial to better predict the ENM fate along with use and end of life, and eventually, develop safer-by-design nano-enabled products.